The Development and Application of a Disaggregate Poisson Model for Trip Generation

摘要

A disaggregate Poisson model for trip generation, prerequisite for modelling travel demand, is presented. The state of the art of modelling trip generation and possible improvements are discussed. In the proposed model, the lambda parameter depends on a number of household characteristics. A maximum likelihood estimation method is discussed of the two cases in which the lambda parameter is additive and multiplicative in the explanatory variables. Examples concerning a submodel of a city transportation study modelling shopping and personal business trips are given. The estimation results appear to be superior to the results of linear and nonlinear regressions on the same data. An analytical aggregation procedure is also discussed.